First-Trimester Placental Growth Factor Shows Modest Value for Predicting Pregnancy Hypertension in Real-World Chinese Cohort

A routine blood test measuring placental growth factor (PlGF) in early pregnancy has long been promoted as a window into placental health, but new research from a Chinese maternity hospital suggests the biomarker performs far less impressively on its own in real-world clinical settings than laboratory studies might imply. In a retrospective cohort study published in Frontiers in Endocrinology, researchers at Hangzhou Fuyang Women And Children Hospital in Zhejiang province traced first-trimester PlGF measurements across hundreds of pregnancy records and found the protein offered only modest discrimination for hypertensive disorders of pregnancy, adding almost nothing beyond basic maternal characteristics such as age, body mass index, and parity.

Hypertensive disorders of pregnancy, which include preeclampsia and gestational hypertension, remain among the leading causes of maternal and perinatal illness worldwide. Preeclampsia is typically defined as new-onset hypertension after 20 weeks of gestation accompanied by proteinuria, maternal organ dysfunction, or uteroplacental dysfunction. Because effective prevention depends on starting aspirin before placental disease becomes clinically established, identifying high-risk women early in pregnancy is a central goal of modern obstetric screening.

PlGF is a proangiogenic member of the vascular endothelial growth factor family, produced by the placenta and circulating in maternal blood. According to the prevailing two-stage placental model, impaired trophoblast invasion and defective spiral-artery remodelling generate placental stress and excess antiangiogenic signalling. Soluble fms-like tyrosine kinase-1, a receptor fragment that binds and neutralises PlGF, rises in the maternal circulation and contributes to the endothelial dysfunction characteristic of preeclampsia. Reduced maternal PlGF may therefore precede clinically apparent disease, although its concentration changes rapidly with gestational age and varies by population and assay platform.

The strongest first-trimester evidence to date concerns multimarker algorithms that combine maternal characteristics, mean arterial pressure, uterine-artery Doppler indices, and biochemical markers. Such composite algorithms outperform risk-factor-only approaches and can identify women who benefit from aspirin prophylaxis. However, translating these findings to Chinese practice is not straightforward, as Asian cohorts show differences in biomarker distributions and screening performance, motivating population- and platform-specific calibration. PlGF testing later in pregnancy among women with suspected preeclampsia serves a different diagnostic purpose and should not be conflated with first-trimester risk prediction.

To examine PlGF’s real-world performance, the team conducted a retrospective, single-centre cohort study using routine-care records from their hospital. The source workbook contained 405 candidate rows, which the researchers carefully deduplicated and audited, resolving 380 unique pregnancy episodes from 380 women. A traceable PlGF value was available for 353 women, and 312 had verified sampling between 9+0 and 13+6 weeks of gestation. After excluding two women with chronic hypertension, the primary cohort comprised 310 women, among whom 27 developed hypertensive disorders of pregnancy: 11 cases of preeclampsia and 16 of gestational hypertension.

PlGF was measured in the hospital laboratory using an automated chemiluminescence platform and reported in picograms per millilitre. Median PlGF was 48.3 pg/mL among women who did not develop hypertensive disorders and 38.1 pg/mL among those who did. When PlGF alone was used to discriminate between outcomes, the area under the curve (AUC) was 0.633, with a 95% bootstrap confidence interval of 0.523 to 0.738. An internally derived PlGF multiple of the median, which standardises values for gestational age, produced a nearly identical AUC of 0.636, indicating little gain from gestational-age adjustment in this cohort.

The researchers then compared PlGF against a parsimonious maternal-characteristics model containing age, pre-pregnancy body mass index, and nulliparity. In 308 complete records, this basic model achieved an apparent AUC of 0.740, considerably better than PlGF alone. Adding log10-transformed PlGF raised the AUC only marginally to 0.748, with a paired difference in AUC of 0.009 and a confidence interval spanning zero. Optimism-corrected AUCs from bootstrap validation were 0.719 and 0.721 respectively. Among the included predictors, pre-pregnancy BMI showed the strongest association with subsequent hypertensive disorders, with an odds ratio of 1.17 per kg/m².

Sensitivity analyses reinforced the pattern. The PlGF AUC was 0.636 when restricted to the narrower 11+0 to 13+6-week window and 0.617 when the two chronic-hypertension episodes were included. A subgroup with documented non-use of aspirin and low-molecular-weight heparin showed an AUC of 0.551, although incomplete medication fields and subgroup selection complicate interpretation. Treating an uncertain gestational-hypertension label as a confirmed event yielded an AUC of 0.619. Across all scenarios, PlGF’s stand-alone discrimination hovered near or below 0.63.

The authors emphasise several important limitations. This was a retrospective single-centre study assembled from available PlGF testing records, with uncertain selection into testing and no hospital-wide denominator. Outcomes were extracted from unvalidated free-text discharge diagnoses rather than independently adjudicated criteria, and gestational age was based on recorded last menstrual period without ultrasound redating. Combining preeclampsia and gestational hypertension may have obscured subtype-specific associations, and the small number of events makes coefficient and calibration estimates unstable despite internal resampling. External validation was not performed.

The findings are consistent with a modest association between lower first-trimester PlGF and subsequent hypertensive disorders of pregnancy, but they do not support isolated clinical use of the biomarker or a deployable prediction model. The results align with the architecture of established first-trimester screening, in which PlGF functions as one component within calibrated multimarker pathways rather than as a stand-alone threshold. Prospective work, the authors suggest, should document the testing pathway, prespecify outcome adjudication, collect mean arterial pressure and uterine-artery Doppler measurements, standardise the assay, and recruit sufficient events to meet calibration-focused sample-size targets before any clinical implementation is considered.

Subject of Research: Placental growth factor screening for hypertensive disorders of pregnancy

Article Title: Real-world discrimination and incremental value of first-trimester placental growth factor for hypertensive disorders of pregnancy: a retrospective Chinese cohort

Article References: Jiang, L., Chai, M., & Xia, Y. (2026). Real-world discrimination and incremental value of first-trimester placental growth factor for hypertensive disorders of pregnancy: a retrospective Chinese cohort. Frontiers in Endocrinology, 17, Article 1960406. https://doi.org/10.3389/fendo.2026.1960406

Image Credits: AI Generated

DOI: 10.3389/fendo.2026.1960406

Keywords: preeclampsia, placental growth factor, hypertensive disorders of pregnancy, first-trimester screening, gestational hypertension, biomarkers, pregnancy, maternal health, Real-world, discrimination, incremental, value